Contribution of nucleoside diphosphokinase to guanine nucleotide regulation of agonist binding to formyl peptide receptors

European Journal ql* Pharmacology - Molecular Pharmacolo,w Section. 208 (1991) 17-23

17

~ 1991 ElsevierScience PublishersB.V. All rights reserved 0922-4106/91/$03.50 ADONIS 0t~2241069100156V EJPMOL 90209

Contribution of nucleoside diphospholdnase to guanine nucleotide regulation of agonist binding to formyi peptide receptors T h o m a s W i e l a n d , J e n s Bremerich, P e t e r Gierschik a n d Karl H. Jakobs Pharmakologisches hzstitut der Unil'er*itfit Heidelberg, D-6900 Heidelberg, F.R.G.

Received i May 1991.accepted 14 May 1991

High-affinity agonist binding to formyl peptide receptors in membranes of myeloid differentiated human leukemia (HL 60) celts is known to be regulated by guanine nuclcotides, most potently by the GTP analog, guanosine-5'-O-(3-thictriphosphate) (GTP[S]). Hcre we analyzed whether nucleosidc diphosphokinasc present in these membranes and capable of formi:~g GTP[S] from GDP and adenosine-5'-O-(3-thiotriphosphate) (ATP[S]) can contribute to nucleotide regulation of agonist receptor binding. Using GDP and ATP[S] at concentrations causing by themselves only small reductions in receptor binding of the labelh:d formyl peptide, N-formyl-methionyl-leucyl-phenylalanine ([3H]FMLP), a marked potentiation (up to 30-fold) was observed when both nucleotides were combined. Under conditions in which the combination of GDP and ATP[S] induced 70-90% of maximal inhibition of [3H]FMLP binding, a total concentration of about 7 nM GTP[S] formed was measured, The synergistic effect of GDP and ATP[S] on [3H]FMLP binding was not seen in the presence '~f UDP (1 mM), which blocked formation of GTP[S] from GDP and ATP[S]. Furthermore, no potentiation was observed when instead of GDP and ATP[S], guanosine-5'-O-(2thiodiphosphate) and adenylyl-5'-imidodiphosphate, respectively, were used. Finally, regulation of [3H]FMLP binding by ATP[S] plus GDP (or GTP) was a time-dependent process, reaching maximal inhibition after 20-30 min of incubation at 25°C. The data indicate that nucleoside diphosphokinase present in membranes of HL 60 cells can nansfcr the thiophosphate group of ATP[S] to GDP leading to formation of GTP[S] and that the GTP[S] thus formed efficiently binds to G proteins interacting with formyl peptide receptors and thereby regulates their agonist binding affinity. Formyl peptide receptor; G protein; Nucleoside diphosphokinase; (HL 60 cells)

1. introduction Formyl peptide receptors in membranes of myeloid differentiated human leukemia (HL 60) cells interact with and activate guanine nucleotide-binding proteins (G proteins) of the G i subtype (Kikuchi et al., 1986, Gierschik and Jakobs, 1987; Gierschik et al., 1989a). Accordingly, agonist binding to these receptors is highly sensitive to regulation by guanine nucleotides. As recently reported (Gierschik et al., 1989b), binding of the chemotactic peptide, N-formyl-methionyl-leucyl-phenylalanine (FMLP), to its receptors in membranes of HL 60 cells is inhibited by both guanine nucleoside tri- and diphosphates. High-affinity agonist receptor binding measured in the presence of Mg -'+ is almost completely reversed or prevented by these nucleotides. Out of the various nucleotides studied, the GTP analog, guanosine-5'-O-(3-thiotriphosphate) (GTP[S]), was

Correspondenceto: K.H. Jakobs, Institutfiir Phare~akologie Universit{it Essen. Hufelandstr. 55, D-4300Essen, F.R.G.

most potent, with half-maximal and maximal inhibition observed at about 10 and 300 nM, respectively. In comparison, GTP and QDP inhibit high-affinity agonist formyl peptide receptor binding with half-maximal -*+ effects at about 1 /xM. In the absence of Mg- , guanine nucleotides aiso inhibited agonist receptor binding, apparently by causing ;~ ;eduction in the receptor affinity for the agonist. Under this condition, GDP inhibited agonist receptor binding with a similar potency as observed in :he presence ,_ff Mg z+, while the potency of GTP[S] was largely reduced and very similar to or even som::what less than that of GDP (Gierschik et at., E-'.8"-)'~,~. As recently reported (Seifert et al., 1988), membranes of differentiated HL 60 cells contain nucleoside diphosphokinase activity. Furthermore, it was shown that the HL 60 membrane nucleoside diphosphokinase can catalyze the formation of GTP[S] from GDP using adenosine-5'-Oo(3-thiotriphosphate) (ATP[S]) as thiophosphate donator. Therefore, we studied whether locally fo;med GTP[S] can contribute to G protein-mediated, guanine nucleotide-induced regulation of agonist

18 binding to formyl peptide receptors. Here, both functional and direct evidence is presented indicating that nucleoside diphosphokinase in membranes of HI, 60 cells is highly efficient in providing GTPiS] from GDP and ATP[S] for G protein-mediated regulation of agonist binding to formyl peptide receptors.

2. Materials and methods

2.1. Materials [SH]FMLP (40-60 Ci/mmol) was flora Nc En gI land Nuclear (Dreieich, F.R.G.) and [8-SH]GDP (10 Ci/mmol) from Amersham Buchler (Braunschweig, F.R.G.). Unlabelled GDP, GTP, GTP[S], ATP[S], guanosine-5'-O-(2-thiodiphosphate) (GDP[S]) ai:d adenylyl-5'-imidodiphosphate (AppNHp), ATP and UDP were from Boehringer-Mannheim (Mannheim, F.R.G.). Unlabelled FMLP was from Sigma (Deisenhofen, F.R.G.). All other materials were from sources previously described -,'Seifet't et al., 1938; Gierschik et al., 1989b). I

2.2. Culture of HL 60 cells and membrane preparation HL 60 cells were grown in suspension culture and were induced to differentiate into mature myeloid cells by cultivation in the presence of 1.25% (v/v) dimethyl sulfoxide for 5 days as described before (Gierschik et al., 1989b). After homogenization of the cells by nitrogen cavitation, crude membranes were prepared by differential centrifugation. The membranes were washed 3 times and finally resuspended to about 10 mg protein/ml and frozen in small aliquots at - 7 0 ° C (Gierschik et al., 1989b). Before assays, membrane aliquots were thawed, diluted with 10 mM triethanolamine-HC1, pH 7.4, containing 5 mM EDTA, centrifuged for 15 min at 30,000×g and 4°C and resuspended in 10 mM triethanolamine-HCl, pH 7.4, at the appropriate membrane concentration.

buffer containing 50 mM Tris-HCl, pH 7.5, 1 mM EDTA and 5 mM MgCI 2. Radioactivity on the dried filters was determined in a scintillation counter. Nonspecific binding was defined as the binding not cempeted for by 10 ~:~M unlabelled FMLP and was less than 5 ~ of total bindi~,g. Data from repeated representative experiments are shown as means of triplicate determinations. The standard deviation of the means was typically less than 5% for all results shown. Protein was determined according to Bradford (1976), using bovine IgG as standard.

2. 4. Formation of GTPIS] b'ormation of GTP[S] from G D P and ATP[S] by HL 60 membranes was determined in a reaction mixture containing 50 mM Tris-HCi, pH 7.5, 1 mM EDTA, 5 mM MgCI z, 0.I ,aM [SH]GDP, 3 ,aM ATP[S] and HL 60 membranes (20 ,ag p r o t e i n / t u b e ) for 30 min at 2 5 ° C in a tota! volume of 50 ,al. The reaction was terminated by addition of perchloric acid (33% (w/v), 5 ,al). After 15 rain at room temperature, neutralization was achieved by addition of 4.3 ,al of 4.5 M KOH containing 100 mM E D T A and 100 mM Tri~-HCI, pH 7.5. After pelleting of denatured proteins, [3H]GTP[S] formed was analyzed by thin layer chromatography as described before (Seifert et al., 19~'3).

3. Results

Binding of [3H]FMLP (2-5 nM) to HL 60 membranes measured in the presence of Mg 2+ (4 mM free concentration) was inhibited by the ATP analog ATP[S]

.-ff

IO[~qb --..--O

I ®'~t. 0.9]. \e

0.6k

2.3. FormylpepHde receptor binding a.,~say

-r

,

O~

o~

' o

coo.o~

\

o

"13 C

Agonist binding to formyl peptide receptors in HL 60 membranes was performed as described before (Gierschik et al., 1989b) in a reaction mixture containing, if not otherwise indicated, 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 5 mM MgCI 2, [~H]FMLP (2-5 nM) and HL 60 membranes (13-23 ,ag protein per tube) in a total volume of 100 ,al. Reactions initiated by addition of membranes to thermally preequilibrated reaction mixtures were performed in triplicate for 30 rain at 25 ° C, The incubation was terminated by rapid filtration through glass fiber filters (Whatman G / C ) and three washes of the filters with 5 ml each of ice-cold

D

o

n

03 GDP

D.-

0

~"

~

lb

16o

ATP[S] (.uM) Fig. l. Potentiation by GDP of ATP[S]-induced inhibition of [3tl]FMLP receptor binding. Bindingof [SH]FMLP(2.7 nM) to HL 60 membranes was determined as described in Materials and methods at the indicated concentrations of ATP[S] in the absence (o, Control) and presence to)of 0.1 # M GDP.

19 (fig. 1). Half-maximal and maximal inhibition (about 80%) was observed at about 30 and 300 /xM ATP[S], respectively. When GDP was additionally present at a concentration (0.1 p.M) causing by itself only a small reduction (5-10%) in [3H]FMLP binding, the potency of ATP[S] ta inhibit formyl peptide receptor binding was largely increased, by a factor of about 30. Halfmaximal inhibition was then obaerved at about 1 g M ATP[S]. At maximally effective concentrations of ATP[S] ( > 100 p.M), no additivit:,, was observed. Vice versa, GDP inhibited [3H]FMLP binding half-maximally and maximally at about 1 and 30 p.M, respectively (fig. 2). In the additional presence of ATP[S] (3 p.M), causing by itself only about a 5% redaction in binding, the potency of GDP to inhibit formyl peptide receptor binding was increased by a factor of about 30. Half-maximal and maximal inhibition by GDP was then observed at about 30 and 300 nM, respectively. When instead of ATP[S] ATP, even at a much higher concentration (100 p.M), was combined with GDP, the potency of GDP to inhibit [3H]FMLP binding was increased by only a factor of about 3 (data not shown). In order to study whether the synergistic effects of GDP and ATP[S] are due to a nucleoside diphosphokinase reaction, transferring the thiophosphate group of ATP[S] to GDP, [3H]FMLP binding and its regulation by GDP and ATP[S] were studied under conditions which are known to inhibit or prevent the thiophosphate transfer reaction. First, in the presence of UDP (1 raM), causing an abortive U D P . nucleoside diphosphokinase complex (Parks and Agarwal, 1973), the potentiating effect of ATP[S] (10 p.M) on inhibition of [3H]FMLP binding by GDP was completely prevented (fig. 3). Similar to ATP, the presence of UDP, which by itself had no effect on [3H]FMLP binding, increased the potency of GDP by a factor of about 3. This effect

.~

1.0r'FW/ I o-',

,

---.,-

,

o~o~

I "\

~ Control

..o

,m.

'-6-"

061

OlGDP

1

~0

{uM)

Fig. 2. Potentiation by ATP[S] of GDP-induced inhibition of [3H]FMLP receptor binding.Bindingof [3H]FMLP (2.1 nM) to HL 60 membranes was determined at the indicated concentrationsof GDP in the absence(o, Control)and presence (o) of 3 ,aM ATP[S],

-~100~'~,...~o~ .@~'~-~--------o~ ~, coo,.o~-] t 't

.

r

\ *so

°

,

!qo \

\

"25: 2

25

A~p

"

0

g'~- obl

c~ ~-~,

~b

(~M) Fig. 3. Influence of UDP on inhibition of [3H]FMLP receptor binding induced by GDF plus ATP[S]. Bindingof [3H]FMLP (2.6 nM) to HL 60 membraneswas determined a~ the indicatedconcentrations of GDP in the absence Co, Control)and presence of either 10 ,uM ATP[S](o), t mM UDP ([3) or the combinationof ATP[S] plus UDP (m). UDP by itself had no effect on {3H]FMLPbinding. both in the absence and presence ol A FP[S].which by itselfreduced control binding(I.04 pmo!/mg pr~+;:in) by 16r~. The ordinalregives [3H]FMLP binding as percentage ~f binding measured in the absenct-of C;DP. GDP

is most likely due to inhibition of degradation of GDP to GMP which is a poor inhibitor or [3H]FMLP receptor binding (Gierschik et al., 1989b). The small inhibition of [3H]FMLP binding induced by 10 # M ATP[S] was not affected by UDP (1 raM) (not shown). Phosphate transfer reactions by nucleoside diphosphokinase require the presence of divalent cations at milfimolar concentrations, with Mg ~+ and Mn ~-+ being rnost effective (Parks and Agarwal, 1973). When [~H]FMLP binding was performed in the absence of Mg 2+, total binding was reduced but inhibition by GDP (0.1 p.M) was preserved and even somewhat larger (on a percentage basis) than in the presence of Mg 2+ as reported before (Gierschik et al., 1989b). ATP[S] at 3 p.M had no effect and at 100 ~M reduced binding by only about 15% (table t). At neither of these ATP[S] concentrations was an additive or synergistic effect with GDP (0.1 # M ) ob,erved. The apparent involvement of a nuc!coside diphosphokinase reaction in the synergistic effect of GDP and ATP[S] on [3H]FMLP binding was, furthermore, corroborated by using alternative nucleoside tri- and diphosphates. When instead of ATP[S] the ATP analog AppNHp (10 IzM) was combined with GDP (0.1 p.M), no potentiation was observed, whereas ATP[S] plus GDP synergistically inhibited [3H]FMLP receptor binding (table 2). Alternatively, when insteaa of GDP the GDP analog GDP[S] was combined w~th ATP[S] (10 p.M), also no potentiation occurr~:d (fig. 4). GDP[S] by itself was 5-10 times more potent than GDP in inhibiting [3H]FMLP receptor binding.

20 -T--//

TABLE l Influence of Mg2" on regulation of [3H]FMLP receptor bind!ng by GDP and ATP[S].

,

\

\ \ ooP:s: o

•-o 50 E D o ..Q

~ 25 u_

903.8 + 38.2 867.1 _+37.0 (4) 283.(/+13.2 (697

763.1 + 33,7 368.3 + 15.7 (52) 221.5+10,0 (71)

No MgCI2 None ATP[S]3/~M ATP[S]IOIT/zM

253.9 -+ 12.0 256.4-+ 82 (01 217. I±10.1(t41

157.5 _+ 6.9 157.5-+ 2.3 (0) 156.5_+ 7.1 (1)

Since the a p p a r e n t t h i o p h o s p h a t e transfer from ATP[S] to G D P is an enzymatic reaction which may require a long period of time to be detectable in the binding assay, time courses of [3H]FMLP binding were performed. A s illustrated in fig. 5, control binding rapidly r e a c h e d equiliorium, within about 10 min at 25°C, and r e m a i n e d constant thereafter. In the presence of G D P ((t.1 / i M ) or A T P (100/~M), equilibrium was apparently o b t a i n e d after 2 - 5 rain with 20% inhibition of binding. At later time points, inhibition of [3H]FMLP r e c e p t o r binding was somewhat d e c r e a s e d , an effect most likely due to d e g r a d a t i o n of these nucleotides. Addition o f ATP[S] (10 # M ) inhibited equilibrium [3H]FMLP r e c e p t o r binding by about 15%. With this nucleotide, equilibrinm was observed already after 2 rain at 25°C, the earliest time point measured,

ob

lb

0:1

guanine nucleotide (uM)

Fig. 4. Influence of ATP[S] on regulation of [3H]FMLP receptor binding by GDP and GDP[S]. Binding of [3H]FMLP (2.6 nM) to HL 60 membranes was determined at the indicated concentrations of either GDP added alone (©) or in combination with 10 ,ttM ATPIS] Co) or GDP[S] added alone (O) or in combination with 10 #M ATP[S} ( I ). ATP[S] by itself reduced control binding (1.04 pmol/mg protein) by 165~. The ordinate gives [3H]FMLP binding as percentage of binding measured in the absence of guanine nucleotides.

W h e n A T P (100 txM) was c o m b i n e d with G D P (0.1 g M ) , the s h a p e of the G D P inhibition curve was not substantially altered, except that the inhibitory effects o f A T P and G D P were roughly additive. C o m p l e t e l y different results, however, w e r e obtait~ed w h e n ATP[S] (10 g M ) was c o m b i n e d with G D P (0.I /xM). A t t h e earliest time point m e a s u r e d (2 min), addition o f ATP[S] to G D P had no or only a small effect. With p r o l o n g e d incubation, [ 3 H ] F M L P b i n d i n g was continuously r e d u c e d by the c o m b i n e d action o f ATP[S] and

Control

"E I "S

-"~A~ p~s!

Gop

E

.... -

11

o.

TABLE 2

~

• ~P:S;

0

0,I #M GDP

5 mM MgCI2 None ATP[S] 3 #M ATP[S] i00 p,M

--

7.=

[ eHJFMLP hound (fmol/mg protein7 Control

,

-6

Binding of [~tI]FMLP t3.7 nM) to HL 60 membranes w::s ~c~crmined as described in Ma'er;als and methods in the absent,: ;. ontrod and presence of 0.1 #M GDP with or without 3 or ",70 gM ATP[S] as indicated. The binding assay was peHormed either in the absence or presence of 5 rrM MgCIz, with 1 mM EDTA being present under each condition Mean values+S,D, of triplicate determinations are given. Numbers in parentheses indicate percent inhibition of binding caused by the addition of ATP[S] at the various conditions. Additions

,

100

~f~

h

GDP° ATP

~...3---- j I

Comparison of the effects of ATP[S] and AppNHp on regulation of [~H]FMLP receptor binding by GDP. Binding of [SH]FMLP t2.5 riM) to HL 6(7 membranes was determined in the absence (Control) and presence of 0.1 #M GDP with and without either 10 #M ATP[S] or 10 t.tM AppNHp. Mean vatues+S.D, of triplicate determinations are given Numbers in parentheses indicate percent inhibitkm of binding caused by the addition of ATP[S] or AppNHp. Additions None ATP[S] 10 g M AppNHpl0gM

[3H]FMLP bound iftool/rag protein) Control

O.1 g M GDP

1 186_+52.9 1043 + 43.(t (12) 1198-+48.7(0)

1 142+_56.1 515 +_21.8 (55) 1106-+44.3(3)

LL

r~

GDP* ATP%:

06

'

3b

incubotion time (rain) Fig. 5, Time course of the synergistic effect of GDP and ATP[S] on []H]FMLP receptor binding, Binding of [3H]FMEP (5.3 nM) to HL 60 membranes was determined for the indicated periods of time at 25°C in the absence to, Control) and presence of either 10 /zM

ATP[S] (o), 100/~M ATP (IS]), 0.1 /z M GDP ( • ) , GDP plus ATP[S] (e) or GDP plus ATP ( I! ).

21 "E 15

Control

/ ,

GDP

~

fi "

~lr

t-

GTP* ATP'S"

~

, -

~

ATPS"

GiP

g -o05 a_

o~

lb

2'o

3'o

incubotion time (rain)

Fig. 6. Time course of the effects of ATP[S]added with either GDP or GTP on [3H]FMLP receptor binding.Bindingof [~H]FMLP (4.8 nM) to HL 60 membraneswas determinedfor the indicated periods of time at 25°C in the absence (o, Control) and presence of either 0.l ,o.MGDP (o), 0.1 /zM GTP (El), 10/xM ATP[S](~,), GDP plus ATP[S](e) or GTP plus ATP[S]( • ).

GDP. A new equilibrium with about 70% inhibition of [3H]FMLP binding was seen after about 20-30 min of incubation at 25°C. Fig. 6 illustrates the time course of the effect of GTP in comparison to GDP (0.1 /zM each) added e~ther alone or in combination with ATP[S] (10 /zM) on [3H]FMLP receptor binding. Both the inhibition curve observed with GTP alone as well as when combined with ATP[S] were very similar to those observed with GDP. At the earliest time point measured (2 min), GTP added alone or in combination with ATP[S] was somewhat more efficient than GDP in inhibiting [3H]FMLP binding, while at later time points the small differences between these guanine nucleotides disappeared. Since all of these functional data strongly suggested that in the presence of ATP[S] and GDP GTP[S] is formed, formation of GTP[S] from GDP and ATP[S] was measured directly using identical assay conditions as for the measurement of [3H]FMLP binding, including membrane concentration (20 ~tg protein) as well as

TABLE 3 Formation of GTP[S] from GDP and ATP[S] in HL 60 membranes. Formation of [3H]GTP[S]from 0.1 ,aM [3H]GDP and 3 ~M ATP[S] was determined in HL 60 membranesas described in Materials and methods in the absence (Control) and presence of 3 nM FMLP with and without 1 mM UDP. Mean values+S.D. of triplicate determinations are given. Additions None UDP 1 mM

[" HjGTP[S]formed (nM) Control 3 nM FMLP 6.57+ 0.20 7.01 + &64 0.09+ t).1)4 0.08 _+0.02

time (30 min) and *emperature (25°C) of the reaction. By the extraction procedure used, total GTP[S] formed, i.e. free plus membrane G protein-bound, was analyzed. With 3 txM ATP[S] and 0.1 txM [3H]GDP as substrates, a condition causing 70-90% of maximal inhibition of [3H]FMLP binding, a total concentration of about 7 nM [3H]GTP[S] was achieved (table 3). The presence of FMLP at a concentration used for the binding experiments shown herein (3 nM) did not cause a significant alteration in formation of [3H]GTP[S]. On the other hand, inclusion of UDP (1 mM) in the reaction almost completely (98-99%) blocked formation of [3H]GTP[S] from ATP[S] and [3H]GDP.

4. Discussion

The data presented herein indicate that membraneassociated nucleoside diphosphokinase can contribute to guanine nucleotide regulation of agonist binding to formyl peptide receptors in membranes of myeloid differentiated HL 60 cells. Using GDP and the thiophosphate analog of ATP, ATP[S], as nucleoside diand triphosphate substrates, respectively, the kinase is apparently capable of transferring the thiophosphate group of ATP[S] to GDP, thereby leading to formation of GTP[S]. This guanine nucleotide then binds with high affinity to membrane G proteins interacting with formyl peptide receptors resulting in reduced agonist receptor binding. The evidence for this chain of reactions is based on the following findings: HL 60 membranes contain nucleoside diphospltokinase enzyme(s), which is capable of converting GDP to GTP[S] using ATP[S] as thiophosphate donator (Seifert et al., 1988; table 3). Second, the synergistic effect of ATP[S] and GDP on [3H]FMLP receptor binding was prevented by UDP, which competes with GDP at the nucleoside diphosphate binding site ef the kinase and at high concentrations leads to an abortive U D P . kinase complex (Parks and Agarwal, 1973). Third, using GDP[S] instead of GDP and AppNHp instead of AT£[S], no potentiation was observed. Compared to GDP, GDP[S] is a pore substrate for nucleoside diphosphokinasc with ATP as nucleoside triphosphate substrate (Goody et al., 1972) and may even be less or not at all thiophosphorylated with ATP[S]. Compared to ATP[S], AppNHp is not a substrate for kinase reactions. Finally, time course exlSeriments revealed that the synergistic action of ATP[S] and GDP on [3H]FMLP receptor binding is a time-dependent process, most likely reflecting the time-dependent, enzymatic formation and accumulation of GTP[S] by the nuc!eoside diphosphokinase. Evidence for nucleoside diphosphokinase-dependem formation of GTP or GTP[S] with subsequent

22 regulation of effector systems known t . be under control of G proteins has been presented for several effector systems, e.g. for stimulation and inhibition of adenylyl cyclase in various tissues, for stimulation of NADPH oxidase in HL 60 cells and for regulation of potassium channels in cardiac membranes (Kimura and Shimada, 1983; Otero et al., 1988; Seifert et al., 1988: Jakobs and Wieland, 1989: Wieland and Jakobs, 1989). To our knowledge, this is the first report describing a regulation of agonist receptor binding by nucleoside diphosphokinase-formed GTP[S]. Several additional aspects have to be discussed here. First, is the GDP transthiophosphorylated to GTP[S] by the nucleoside diphosphokinase free or bound to G proteins? The data reported herein do not allow a firm conclusion. The only argument in favor of G proteinbound GDP being the substrate for the kinase could be the action of ATP[S] itself. Compared to ATP inhibiting [3H]FMLP binding half-maximally at _>1 mM (Gierschik et al., t989b), ATP[S] was clearly more potent with half-maximal inhibition occurring at about 30 ~zM. However, a similar large difference in potencies is observed comparing GTP and GTP[S] (Gierschik et al., 1989b). Furthermore, in the absence of Mg 2+, i.e. without functional r~ucleoside diphosphokinase, GDP inhibited [3H]FMLP receptor binding at least as well as in the presence of Mg 2+, whereas ATP[S] was far less potent in the absence than in the presence of Mg 2+. However, similar differences were obtained for GTP[S], inhibiting [3H]FMLP binding half-maximally at about 10 nM and 1 g M in the presence and absence of Mg 2+, respectively (Gierschik et al., 1989b). Accordingly, the fact that in the absence of Mg -'+ no synergistic effect of ATP[S] and GDP on [3H]FMLP receptor binding was observed is not only due to an inactive nucleoside diphosphokinase enzyme but also a largely reduced potency of possibly formed GTP[S] under this condition. It is conceivable that ATP[S] binds to G proteins in a manner similar to GTP[S] but with a largely reduced affinity (at least 3 orders of magnitude) or that ATP[S] is contaminated by a small amount (less than 0.1%) of GTP[S], being finally responsible for the effects observed with ATP[S] when added alone. Thus, althougn involvement of a nucleoside diphosphokinase reaction in the effect of ATP[S] alone cannot be excluded, the alternative interpretations arc also not unlikely. Furthermore, the membrane preparation may contain loosely bound GDP to which the kinase has access. Thus, in order to answer the question whether G protein-bound GDP can be transthiophosphorylated with ATP[S] to GTP[S] by nucleoside diphosphokinase, studies have to be performed with components free of any contaminants and with GDP tightly bound to G proteins. The second aspect which has to be discussed here is the question whether GTP[S] (or GTP) locally formed

by the nucleoside diphosphokinase has any advantage in activating G proteins over GTP[S] (or GTP) exogenously added to the system. The data reported herein do not exclude this possibility and may even argue in favor of a higher efficiency of locally formed GTP[S] than of added GTP[S]. For example, under a condition in which the combined action of GDP and ATP[S] induced 70-90% of the maximal effect of nucleotides on [~H]FMLP receptor binding, a total concentration of about 7 nM GTP[S] formed was measured. On the other hand, GTP[S] exogenously added to the system caused 50% of its maximal effect on [3H]FMLP binding at a concentration of about 10 nM (Gierschik et al., 1989b). These data, thus, seem to suggest that locally formed GTP[S] can be more efficient (about 2- to 3-fold) in activating G proteins than GTP[S] being exogenously added and uniformly distributed in the incubation medium. In :,rder to corroborate such a hypothesis, studies are in progress to analyze G protein regulation in HL 60 membranes by nucleoside diphosphokinase-formed GTP[S] and exogenously added GTP[S] under a variety of conditions, by measuring both G protein function (regulation of agonist-receptor binding, receptor-stimulated GTPase activity and GTP[S] binding) and GTP[S] formation under identical conditions. Finally, one has to ask whether such transphosphorylation reactions by nucleoside diphosphokinase at the plasma membrane, capable of delivering GTP to G proteins, are physiologically relevant for G protein function in intact cells, particularly considering the higher total cellular concentration of GTP compared to that of GDP (Kleineke et al., 1979; Sklar et al., 1989). However, at the plasma membrane, the hydrolysis of GTP and, thus, the formation and accumulation of GDP, which also arises from release from G proteins, can be locally high. The nucleoside diphosphokinase enzymes will transphosphorylate GDP formed from GTP hydrolysis back to GTP as well as direc~'ly added GDP, as shown herein for apparent equal formation of GTP[S] from either GDP or GTP in the presence of ATP[S] (fig. 6). The questioq whether such a transphosphorylation reaction is required for or facilitates G protein activation can only be answered by affecting the nucleoside diphosphokinase enzymes. However, no specific, membrane-permeable inhibitors of these enzymes are known. Recently, several nucleoside diphosphokinase enzymes have been cloned (Kimura et al., 1990; Lacombe et al., 1990; Mufioz-Dorado et al., 1990) and, most interestingly, and shown to be very homologous or even identical to Nm23, a protein apparently involved in mammalian tumor metastasis (Liotta and Steeg, 1990; Wallet etal., 1990). It will be interesting to learn whether over-expression or inhibition of expression of these kinases affects signal transduction via G proteins in intact cells.

23 Acknowledgements We are indebted to Marlene Beck and Susanne Gierschik for expert technical assistance. This work was supported by the Deutsche Forsch ungsgemeinscha ft.

References Bradford, M.M., 1976, A rapid and sensitive method for the quantiration of microgram quantities of protein utilizing the principie of protein dye binding, Anal. Bio':hem. 72, 248. Gierschik, P. and K.H. Jakobs, 1987, Receptor-mediated ADP-ribosylation cf a phospholipase C-stimulating G protein, FEBS Lett. 224, 219. Gier~chik, P., D. Sidiropoulos and K.H. Jakobs, 1989a, Two distinct Gi-proteins me, iate formyl peptide receptor signal transduction in human leukemia (HL-60) cells, J. Biol. Chem. 264, 21470. Gierschik, P., M. Steisstinger, D. Sidiropoulos, E. Herrmann and K.H. Jakobs, 1989b, Dual Mg a+ control of formyl-peptide-receptor-G-protein interaction in HL 60 cells. Evidence that the !owagonist-affinity receptor interacts with and activates the G-protein, Eur. J. Biochem. 183, 97. Goody, R.S., F. Eckstein and R.H. Schirmer, 1972, The enzymatic synthesis of thiophosphate analogs of nucleotides, Biochim. Biophys. Acta 276, 155. Jakobs, K.H. and T. Wieland, 1989, Evidence for receptor-regulated phosphotransfer reactions involved in activation of the adenylate cyclase inhibitory G protein in human platelet membranes, Eur. J. Biochem. 183, 115. Kikuchi, A., O. Kozawa, K. Kaibuchi, T. Katada, M. Ui and Y. Takai, 1986, Direct evidence for involvement of a guanine nuclcotide-binding protein in chemotactic peptide-stimulated formation of inositol bisphosphate and trisphosphate in differentiated human leukemic (HL-60) cells. Reconstitution with G i o r G~j of the plasma membranes ADP-ribosylated by pertussis toxin, J Biol. Chem. 261, 11558. Kimura, N. and N. Shimada, 1983, GDP does not mcdiate but rather inhibits hormonal signal to adenylate cyclasc, J. Biol. Chem. 258, 2278.

Kimura, N., N. Shimada, K. Nomura and K. Watanabe, 1990, Isolation and characterization of a eDNA clone encoding rat nucleoside diphosphate kinase, J. Biol. Chem. 265, 15744, Kleineke, J., C. Diils and H.-D. SOling, 1979, Subcellular compartmentation of guanine nucleotides and functional relationships between the adenine and guanine nucleotide systems in isolated hepatocytes, FEBS Left. 107, 198. Lacombe, M.-L.. V. Wallet, H. Troll and M. V&on, 1990. Functional cloning of a nucleoside diphosphate kinase from Dictyostelium discoideum, J. Biol. Chem. 265, 10012. Liotta, L.A. and P.S. Steeg, 1990, Clues to the function of Nm23 and Awd proteins in development, signal transduction, and tumor metastasis provided by studies of Dictyostelium discoideum, J, Natl. Cancer Inst. 82, 1170. Mufioz-Dorado, J,, M. Inouye and S. Inouye, I990, Nuclcoside diphosphate kinase from Myxococcus xanthus. I. Cloning and sequencing of the gone, J. Biol. Chem. 265, 2702. Olero, A.S., G.E. Breitwieser and G. Szabo, 1988, Activation of muscarinic potassium currents by ATP3,S in atrial cells, Science 242, 443. Parks, R.E. and R.P. Agarwal, 1973, Nacleoside diphosphokinases, in: The Enzymes, Vol. 8, ed. P.D. Boyer (Academic Press, New York) p. 307. Seifert. R., W. Rosenthal, G. Schultz, T. Wieland, P. Gierschik and K.tt. Jakobs, 1988, The ro',e of nucleoside-diphosphate kinase reactions in G protein activation of NADPH oxidase by guanine and adenine nucleotides, Eur. J. Biochena, 175, 51. Sklar, L.A., H. Mueller, G. Omann and Z. Oades, 1989, Three states for the formyl peptide receptor on intact cells. J. Biol. Chem. 264, 8483. Wallet, V., R. Mutzel, H. Troll, O. Barzu, B. Wurster, M. Vdron and M.-L. Lacombe, 199(t, Dictyostelium nucleoside diphosphate kinase highly homologous to Nm23 and Awd proteins involved in mammalian tumor metastasis and Drosophila development, J. Natl. Cancer Inst. 82, 1199. Wieland, T. and K.H. Jakobs, 1989, Receptor-regulated formation of GTP[),S] with subsequent persistent G,-prolein activation in membranes of human platelcts, FEBS Left. 245, 189.